Skip to main content

Table 3 Average error of heterogeneity estimation

From: SHEAR: sample heterogeneity estimation and assembly by reference

  Deletions Tandem duplications
Depth 150 bp 1000 bp 30 kbp 150 bp 1000 bp 30 kbp
10 × 0.00 1.22 1.46 13.98 13.73 12.27
20 × 0.08 0.52 1.14 8.91 11.22 12.40
30 × 0.17 0.42 1.06 9.64 11.48 11.86
50 × 0.10 0.27 1.17 9.83 9.69 9.20
100 × 0.08 0.21 1.21 5.56 5.47 6.06
500 × 0.09 0.21 1.04 4.17 2.35 4.09
1000 × 0.08 0.18 1.06 4.58 2.11 3.40
  1. Each entry reports the average absolute error for estimation of heterogeneity percentage for a variety of SVs at different overall coverage levels. For each pairing of SV type and coverage level, ten iterations of simulation were sampled from each of seven different underlying heterogeneity percentages (20%, 40%, 60%, 80%, 90%, and 100%) for a total of 70 simulations per entry in the table. The reported error is the absolute difference between SHEAR’s estimation of heterogeneity percentage and the true percentage of breakpoint reads originating from the variant sequence. Each entry in the table contains the average error for that scenario, ignoring simulations in which the SV was not predicted. For example, for the first entry in the table (150 bp deletion at 10 × depth), the SV is only predicted in 24 out of the 70 simulations due to the low coverage, and thus the average error is from those 24 estimations.